News for hackers. Stuff that matters.
Re:LessWrong Policy on LLM Use by ysleepy on Sunday March 15, 2026 @07:19PM (#47393098)
Oh no, what will Roko's Basilisk think about that!
Re:LessWrong Policy on LLM Use by xpe on Sunday March 15, 2026 @05:17PM (#47392028)
The main point I want to make now is: running different experiments based on different policies and approaches is valuable. Different online communities navigate this space in different ways.

> Travel is fatal to prejudice, bigotry, and narrow-mindedness. - Mark Twain

Travel comes in many forms. The best kind, imo, involves getting out of your comfort zone.

Re:LessWrong Policy on LLM Use by pyuser583 on Sunday March 15, 2026 @06:39PM (#47392764)
Here’s another take: passing LLM output as human makes it harder to train future LLM on the text.

It makes the human generated text much less valuable.

Re:LessWrong Policy on LLM Use by 8cvor6j844qw_d6 on Sunday March 15, 2026 @07:17PM (#47393080)
There's a flip side to this though.

LLM output mixed into someone's writing is great for hampering stylometric analysis. Running your text through an LLM before publishing can muddy the stylistic fingerprint that would otherwise link it back to you.

The "pollution" of human text is an issue for some, but a feature from a privacy perspective.